4 ms·
No, it’s not similar (other than superficially, as both use diffusion methods in some way). It uses diffusion to generate latent vectors that are entered in a n
by woodson 4y ago
No, it’s not similar (other than superficially, as both use diffusion methods in some way). It uses diffusion to generate latent vectors that are entered in a neural audio codec model to produce speech.